Personality is impossible to define succinctly because it means different things to different psychologists. Whilst most would accept that the field of personality is the study of how individuals differ from one another, they would disagree on the best way to conceptualise these individual differences. One definition is given by Allport (1961) who suggests that personality is ‘a dynamic organisation, inside the person, that creates the person’s characteristic patterns of behaviour, thoughts and feelings’. On the other hand personality traits or dimensions (e.g. extraversion) are relatively stable dispositions that give rise to characteristic patterns of behaviour. This essay will outline and evaluate the five factor approaches (five factor model and the big five) to personality, looking at theories put forward by Tupes and Christal (1961), Goldberg and Costa and McCrae.
Psychologists increasingly agree that five supertraits may adequately describe the structure of personality. The evidence to support this contention has come from a number of sources. There is still some debate about how to label these factors but this is perhaps unsurprising given that assigning labels is the most subjective aspect of factor analysis. Researchers are likely to have different opinions about which words describe the consistent traits that make up a supertrait. This essay shall begin by examining the evidence for five factors. Evidence for the five factor model comes partly from the lexical approach and partly from factor analysis.
According to the lexical approach it is the differences in personality that are important for social interaction, and human societies have labelled these differences as single terms. Cattell’s 16 PF came from the factor analysis of the lost of 4,500 trait names identified by Allport and Odbent (1936). Cattell than reproduced a 16 factor solution. Fiske (1949) reanalysed the same data but could not reproduce the 16 factors; he published instead a five factor solution. This work was ignored for a long time. Tupes and Christal (1992) reported five factors from analyses of trait words in 8 different samples. Norman (1963) revisited the earlier research and reproduced the same five factor structure using personality ratings of individuals given by their peers. Digman and Takemoto-chock (1981) carried out further analysis and confirmed Norman’s five factor solution.
Goldberg (1981) reviewed all the research and made a convincing argument for the big five. Goldberg (1990) concluded that in the English language trait descriptors are versions of five major features of personality: love, work, affect, power and intellect. Since then, the research has spread to other languages. Saucier and Ostendorf (1999) used a set of five hundred personality traits and found a five factor structure in the German language, for example.
Saucier and Goldberg (2001) have described the lexical approach to investigating whether the five factor structure is universally applicable as an emic approach to research. Basically what the researchers do is use the personality terms that are found in the native language of the country. They contrast this with what they call the etic approach, which uses personality questionnaires translated from another language that in practice tends to be English. Saucier and Goldberg (2001) report that etic approaches tends to replicate the five factor structure while there is more variability reported in studies using emic approaches. Perugini and Di-Blas (2002) discuss this issue further in relation to emic and etic data they collected on Italian samples. They point out that in the etic approach; the questionnaires being translated are based on five factor structures found in the original language.
Evidence for the existence of five dimensions of personality also comes form factor analysis. Costa and McCrae (1997) are arguably the most influential researchers in this area, and their factor solution has come to be called the big five model. This approach requires large samples of participants to complete at least two personality questionnaires. The resultant data set is then factor-analysed to uncover clusters of traits. The consistent finding is the emergence of five factors or dimensions of personality.
It is important to stress that it is the analysis of data that has produced the factors, not exploration of a theory about the number of factors necessary in a model. This is not the usual approach in Psychology. Usually researchers begin with a theoretically based hypothesis about some aspect of behaviour. They then collect their data, and their results either support or refute their original theory driven hypothesis. In contrast, with the five-factor research, the hypothesis that five factors represent the basic structure of personality has come from the data that was collected. In other words the big five is a data derived hypothesis as opposed to a theoretically based one.
These are the factors described by Costa and Mc Crae (1992), who measured personality with their well known personality inventory (NEO-PIR). Each factor represents a continuum along which individuals can be placed according to their scores.
These are the five main dimensions popularly known as the big five. Within each of the main dimensions there are more specific personality attributes that cluster together, and all contribute to the category score. These subordinate traits are sometimes called facets (Costa and McCrae 1992). The big five model is a hierarchical model similar in concept to Eysenck’s model. Each of the big five factors consists of six facets or subordinate traits .Thus for example, an individuals scores on the traits of fantasy, aesthetics, feelings, actions, ideas and values combine to produce their scores on the openness factor. The NEO-PIR then allows measurement at a general factor level or on more specific factors. Obviously, the more specific the measure, the greater the likelihood of using it to actually predict behaviour.
Moreover, In terms of how well this model fits with other measures of personality, the evidence is largely positive. McCrae and Costa (1989) factor-analysed scores on the Myers-Briggs Type Inventory and found that it supports a five factor structure. Boyle (1989) reported that the five factor model is also broadly compatible with Cattell’s 14 factor measure and Eysenck’s three factor measure. The latest measure of the 16PFI allows scoring on the big five (Conn and Reike 1994). Goldberg (1993) compared the five factor model with Eysenck’s three factor model and concluded that two of the factors – extraversion and neuroticism – are very similar, and that psychoticism can be subsumed under agreeableness and conscientiousness.
The NEO-PIR has also been translated into several other languages, and the same factor structure has been replicated (McCrae and Costa 1997). This evidence is no uncontentious, based as it is on the etic approach to personality research. McCrae and Costa (1997) also demonstrated that the observed personality differences are stable over time and have a genetic basis. To summarise Costa and McCrae (1992) claim that the five factors represent the universal structure of personality based on all the evidence discussed in this essay. These factors are found in different languages, ages of people and races.
Another major advantage of the big five framework is that it can assimilate other structures. For example, Goldberg and Rosolack (1994) demonstrated empirically that Eysenck’s three –factor system of extraversion, neuroticism and psychoticism can be integrated into the big five. Psychoticism is a combination of undesirable big five III (low conscientiousness) and undesirable big five II (low agreeableness), Eysenck‘s extraversion is equivalent to big five I (also called extraversion) and Eysenck’s neuroticism to big five IV (emotional stability, which is simply the desirable pole of neuroticism). The big five makes a useful structure for organising the large and confusing number of traits and their measures in vogue today. However it remains, for the most part, a description of normal personality and therefore is not as useful in clinical applications as it is in other areas such as occupational psychology.
In light of this can we then conclude that the big five represent the structure of personality? Unfortunately, it is premature to say that there is total consensus on the model. There is an increasing agreement that there are five factors, but there is still some level of disagreement about the exact nature of each of these factors. Indeed, Saucier and Goldberg (1998) argue that research should look for solutions beyond the current five factor models. This is the scientific approach – to search for contradictory evidence instead of purely focusing on searching for confirmation, as the present research does.
Peabody and Goldberg (1989) have also demonstrated that the measures that are included in a questionnaire crucially affect the final factors produced. If a questionnaire does not have any items that measure openness, for example, than the description of openness that is produced will be narrower. There is still some argument about the number of traits, with studies reporting different numbers between Eysenck’s three and seven (Briggs 1989). McCrae and Costa (1995) suggest that the number depends on the nature of the trait measures that are included. They point out that five factor models tend not to include evaluative traits like moral/immoral. If evaluative traits are included, Almagore et al (1995) have suggested that a seven factor solution emerges.
There has also been some debate as to what the factors mean (Digman 1990). Are they perhaps linguistic categories that perhaps do not represent the underlying structure of personality? Is it that the five factors represent our ability to describe personality traits in language and are nothing to do with underlying structures? There is no easy answer to this question, although the accumulating evidence would seem to negate it. Is it perhaps that out cognitive abilities only allow for a five factor structure but the reality is more complex and subtle?
Briggs (1989) has criticised the model for being atheoretical. The model is data driven and was not derived from a theoretical base. There are currently some attempts to address this with genetic studies and the search for a physiological basis for the observed differences. This criticism applies more generally to the trait approach, although theorists such as Eysenck saw theory building as crucial within his approach.
The study of the biological basis of personality can also provide support for the five factor approach, a number of biological ‘causes’ have been postulated for the existence of different traits, some of which are identified by the five factor approaches. A problem with this however is that many of these biological ‘causes’ are postulated at “conceptual nervous system” level, for example unrealistic and oversimplified notions of arousal are applied. The best known example is that of Eysenck’s accounts of extraversion and neuroticism. Eysenck’s concept of extraversion (equivalent to big five I – also called extraversion) has been associated with cortical arousal via the ARAS (Ascending reticular activating system).
According to the Yerkes-Dodson law people have an optimal arousal level of performance in learning tasks, if arousal levels are low, performance is also low, for example boredom leads to lack of attention and this in turn leads to lowered performance. High arousal can also lead to poor performance, for example a death in the family. So, the law suggests that people need an equilibrium/balance for optimal performance. This is linked to personality because extraverts are said to be chronically under-aroused and seek to increase their arousal through behavioural means, whilst introverts are chronically over aroused and seek to decrease their arousal through behavioural means. Evidence for this theory comes from a number of sources. Firstly people who tend to score low on extraversion also tend to engage in quieter pursuits. So their behaviour matches their score on personality tests measuring for introversion and extraversion.
Moreover Gale (1983) carried out EEG measurement of cortical arousal and found that introverts were more aroused in 22/38 comparisons. A higher EEG reading for introverts than extroverts suggests that there are real physiological differences in brain activity between introverts and extroverts hence supporting the existence of at least one of the traits identified by the five factor approaches.
Another trait identified by five factor approaches is also identified physiologically in the human body. Neuroticism (equivalent to big five IV), has been related to the sensitivity of the limbic system which affects the autonomic nervous system. The theory is that the higher the person scores on neuroticism the more physiological symptoms they should have (for example sweating, respiration, rapid heart rate etc.)
This is because their limbic system (including the hippocampus and hypothalamus) is more sensitive and so they are more likely to react by acting on the sympathetic branch (flight or flight rather than rest and digest). Higher scorers of neuroticism do report more physiological symptoms but the problem is that these are not always measurable and often here are different patterns of responses (response specifity). Moreover there are no clear differences between high and low neuroticism scorers.
Gray’s two factor (reinforcement sensitivity) model provides a clearer physiological basis which is more developed from Eysenck’s work. Gray’s model is rooted in animal learning research and revolves around two brain systems – Behavioural approach system (BAS) and behavioural inhibition system (BIS). According to the model BAS is the brain system that makes a person sensitive to reward. More sensitive individuals display more approach (reward seeking) behaviours; impulsivity.
BIS on the other hand is the brain system that is sensitive to punishment. More sensitive individuals hence show more avoidant and anxious behaviour. Not only have neuroticism and extroversion been shown to exists as real traits, so has Psychoticism (also known as undesirable Big five III). For example those with higher testosterone levels are those who usually score higher on Psychoticism measures in personality tests.
One of the more general criticisms of the trait approaches to personality is related to how the various measures are interpreted and used. For example Mischel (1990) has pointed out that many of these measures are largely descriptive and do not predict behaviour particularly well. Despite this claim, many of these measures are widely used to make important decisions about individuals’ lives and in workplace situations.
They are also widely interpreted by people who are not psychologists. Mischel (1968) demonstrates that on average, personality trait measures statistically account for only around 10% of variance observed in behaviour. In other words 90% of behaviour is down to factors other than personality. However Kraus (1995) has shown that the variance figure is not insignificant and is similar to that found in studies measuring the relationship between attitudes and behaviour.
When considering the nature-nurture debate most research supports the idea that behaviour is the result of a complex interplay of traits and situations (e.g. Plomin 1994). The big five exclude the fact that the situation may have an effect on the way a person behaves and puts way too much emphasis on innate traits. This is another criticism that can be made of trait theories such as big five.
Mischel (1968) compiled a critique of traits in which he demolished the idea that people behave consistently regardless of the situation. Winter et al (1998) supported this by arguing that motives describe unique aspects of personality not captured by traits, and that both units of analysis are necessary for a compete understanding of personality.
Despite the alternatives to trait theories based on criticisms (such as theories based on motives and cognition), a significant chunk of personality psychology is still built around the concept of traits. The big debate of the 1990’s has been about the structure of trait terms; in particular, how many broad traits are needed to provide a comprehensive description of personality. Eysenck (1991) remained a strong advocate of three: extraversions, neuroticism and psychoticism, whereas Cattell believed in about fifteen plus intelligence. However, the winning number in this lottery is five. Digman (1990), Goldberg (1993), John (1990) and McCrae and Costa (1997) are all compelling advocates for a five factor structure, composed of broad domains of personality known as the big five.